Low cost storage for rarely read data

• US 9,841,774 B2
• Filed: 10/17/2016
• Issued: 12/12/2017
• Est. Priority Date: 05/21/2013
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Balakrishnan, Shobana M., Harper, David T., Heil, Stephen, Peterson, Eric C., Glass, Adam B., Butler, David Alex, Donnelly, Austin Nicholas, Rowstron, Antony Ian Taylor, Legtchenko, Sergey
• Primary Examiner(s): SHIN, CHRISTOPHER B
• Application Number:
  US15/295,971
• Publication Number:
  US 20170031372A1
• Time in Patent Office: 421 Days
• Field of Search: 713323-324, 710 74, 711114
• CPC Class Codes: G05D 23/1917 using digital means
• CPC Class Codes: G06F 1/20 Cooling means
• CPC Class Codes: G06F 1/26 Power supply means, e.g. regulation thereof for memories G11C
• CPC Class Codes: G06F 13/4068 Electrical coupling
• CPC Class Codes: G06F 3/0611 in relation to response time
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0625 Power saving in storage systems
• CPC Class Codes: G06F 3/0634 by changing the state or mode of one or more devices
• CPC Class Codes: G06F 3/0658 Controller construction arrangements
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Low cost storage for write once read rarely data is described. In an embodiment a storage device comprises a plurality of hard disk drives connected to a server via an interconnect fabric. The storage device comprises a cooling system which is only capable of cooling a first subset of the hard disk drives and a power supply system which is only capable of powering a second subset of the hard disk drives and in some examples, the interconnect fabric may be only capable of providing full bandwidth for a third subset of the hard disk drives. Each subset may comprise only a small fraction of hard disk drives. A control mechanism, which may be implemented in software, is provided which controls which hard disk drives are active at any time in order that the constraints set by the cooling and power supply systems and interconnect fabric are not violated.

Claims

1. An electronic storage system comprising:

a plurality of computer storage media, each of the plurality of computer storage media having an active state and a non active state;

a power supply system capable of providing sufficient power for only a subset of the plurality of computer storage media to be in an active state concurrently;

a cooling system provisioned to provide sufficient cooling when operating for only a subset of the plurality of computer storage media to be in an active state concurrently;

a control mechanism arranged to dynamically control which of the plurality of computer storage media are in an active state; and

at least one of;

one or more of the plurality of computer storage media having a maximum bandwidth in the active state;

at least a portion of the plurality of computer storage media being arranged logically into groups, each group comprising a plurality of computer storage media capable of being in an active state concurrently;

or

each computer storage media belonging to at least one of a cooling domain and a power domain, a cooling domain comprising one or more computer storage media linked by a cooling constraint and a power domain comprising one or more computer storage media linked by a power constraint.

2. The electronic storage system according to claim 1, wherein the plurality of computer storage media are located within a device.

3. The electronic storage system according to claim 1, wherein the control mechanism is arranged to dynamically control which of the plurality of computer storage media are in an active state using the maximum bandwidth of at least one of the one or more the plurality of computer storage media having a maximum bandwidth in the active state.

4. The electronic storage system according to claim 1, further comprising an interconnect fabric connecting the plurality of computer storage media and a server, the interconnect fabric comprising a plurality of SATA controllers and PCIe switches, wherein each of the plurality of computer storage media is connected to the server via a SATA controller and one or more PCIe switches arranged in a tree structure and wherein the PCIe switches are physically distributed within the electronic storage system.

5. The electronic storage system according to claim 1, further comprising an interconnect fabric connecting the plurality of devices and a server, the interconnect fabric comprising a plurality of SATA multiplexers arranged in a tree structure.

6. The electronic storage system according to claim 1, further comprising a server comprising a data store arranged to store a ‘

no access’

flag associated with each of the plurality of computer storage media and when set, a ‘

no access’

flag causes all IO requests on a computer storage media to fail.

7. The electronic storage system according to claim 1, wherein at least one of the plurality of computer storage media includes a spinning disk in an active state.

8. The electronic storage system according to claim 1, wherein the cooling constraint corresponds to characteristics of the cooling system and the power constraint corresponds to characteristics of the power supply system and wherein each group comprises a plurality of computer storage media in non-overlapping cooling domains and non-overlapping power domains.

9. The electronic storage system according to claim 8, wherein the plurality of computer storage media are arranged logically into groups such that each group comprises a plurality of computer storage media in non-overlapping cooling domains and non-overlapping power domains and wherein groups are arranged to be either fully colliding or disjoint, wherein two groups are fully colliding if each computer storage media in a first group is a member of the same cooling and power domain as computer storage media in a second group and two groups are disjoint if each computer storage media in a first group is not a member of the same cooling or power domain as any computer storage media in a second group.

10. The electronic storage system according to claim 1, further comprising a server comprising a scheduler arranged to:

receive a burst of data to be written to the electronic storage system divided into portions;

add error correction data to each portion; and

write each portion to computer storage media from a single group.

11. The electronic storage system according to claim 1, wherein each portion is divided into a plurality of stripes and each stripe is divided into a plurality of blocks, and wherein adding error correction to each portion comprises:

adding one or more blocks comprising redundancy information to each stripe and wherein writing each portion to computer storage media from a single group comprises;

for each stripe, writing one block from the stripe to a different one of the computer storage media from the single group.

12. The electronic storage system according to claim 11, wherein for each stripe, writing one block from the stripe to a different one of the computer storage media from the single group comprises:

assembling a sequence of blocks comprising one block from each stripe; and

writing each sequence of blocks to a separate computer storage media from the single group.

13. The electronic storage system according to claim 1, further comprising a server comprising a scheduler arranged to:

reorder a plurality of operations into sets of operations operating on the same group of computer storage media, the operations comprising one or more of read, write and delete operations; and

schedule sets of operations in an order which maximizes throughput of the electronic storage system.

14. The electronic storage system according to claim 13, wherein the scheduler is further arranged to receive a burst of operations and to perform reordering on the burst of operations.

15. The electronic storage system according to claim 13, wherein the scheduler is arranged to perform reordering on a subset of operations in a queue of operations, wherein the subset of operations is defined by a window and wherein the window has a length specified in terms of a number of operations or a period of time.

16. The electronic storage system according to claim 13, wherein the scheduler is arranged to schedule sets of operations in an order which minimizes switching between groups of computer storage media.

17. The electronic storage system according to claim 1, further comprising a server comprising a scheduler arranged to:

receive a burst of data to be written to the electronic storage system divided into portions;

divide each portion into a plurality of segments;

add one or more error correction segments; and

write each segment to a different computer storage media from a single group.

18. The electronic storage system according to claim 1, wherein each computer storage media belongs to a cooling domain and a power domain, a cooling domain comprising computer storage media linked by a cooling constraint and a power domain comprising computer storage media linked by a power constraint, the cooling constraint corresponding to characteristics of the cooling system and the power constraint corresponding to characteristics of the power supply system,

the electronic storage system further comprising a server comprising a scheduler arranged to:

identify a subset of the computer storage media storing data corresponding to a read request;

determine a migration path from a current configuration comprising those computer storage media in an active state to a target configuration comprising the identified subset of computer storage media in active state via a plurality of intermediate configurations, wherein each configuration satisfies all power and cooling constraints; and

migrating the current configuration of the electronic storage system to the target configuration via the plurality of intermediate configurations.

19. A method of operating an electronic storage system comprising a plurality of computer storage media and insufficient cooling or power infrastructure for all of the computer storage media to be in an active state concurrently, wherein at least a portion of the plurality of computer storage media are arranged logically into groups, each group comprising a plurality of computer storage media capable of being in the active state concurrently, the method comprising:

reordering a plurality of operations into sets of operations operating on the same group of computer storage media; and

scheduling sets of operations in an order which maximizes throughput of the electronic storage system.

20. An electronic storage system comprising:

a plurality of computer storage media, each computer storage media having an active state and a non active state, and at least a portion of the plurality of computer storage media being arranged logically into groups, each group comprising a plurality of computer storage media capable of being in an active state concurrently while complying with at least one of a bandwidth constraint, a power constraint, or a cooling constraint;

at least a portion of the computer storage media belonging to at least one of a cooling domain or a power domain, a cooling domain comprising computer storage media linked by a cooling constraint and a power domain comprising computer storage media linked by a power constraint;

at least one of;

a power supply system capable of providing sufficient power for only a subset of the computer storage media to be in an active state concurrently;

or

a cooling system provisioned to provide sufficient cooling when operating for only a subset of the computer storage media to be in an active state concurrently; and

a control mechanism arranged to dynamically control which groups of computer storage media are in an active state according to a request received by the server.